190 ON THE COLOURS OF THICK PLATES.
of the light by which the rings are formed is confirmed by the results of several experiments.
In the experiments of Sir William Herschel and M. Pouillet mentioned in the introduction, as well as in some of those of the Duke de Chaulnes, rings of the same nature as those formed by a tarnished mirror of quicksilvered glass were produced in cases in which the deflection of the light from its regular course was in-contestably of the nature of diffraction. From the similarity of effect we have a right to infer a probable similarity of cause, unless such a supposition should entail some peculiar difficulty, which does not seem to be the case in the present instance, but quite the contrary.
31. Having cleaned the surface of a concave mirror which had been employed in forming the bands, I rubbed a little tallow on it, and then wiped the mirror in one direction with a handkerchief, so as to have a finely striated tarnish on it. The tarnish was not sufficient to cause much obscurity; but the image of a candle seen in the mirror was accompanied, as is usual in such cases, by two tails of light, which ran out in a direction perpendicular to the striae. Having placed a small flame in the centre of curvature of the mirror, I found that the rings were formed with great brilliancy where they were intersected by the tails of light, but elsewhere they were almost wholly invisible.
Now the tails of light are known to be a phenomenon of diffraction: the striated tarnish may in fact be regarded as a sort of irregular grating, and the tails of light are of the nature of Fraunhofer's spectra. If a tarnish in general were capable of producing rings independently of diffraction, there appears no reason why a tarnish of tallow should not be capable; for the particles of tallow are as fine as those of most other kinds of tarnish. But if in the case of a tarnish of tallow the deflection of the light from its regular course be not a phenomenon of diffraction, there appears no reason why the rings should be confined to the tails of light in the experiment described above.
32. The phenomena of polarization seem however to lead to a crucial experiment for deciding whether the deflection of the light from its regular course, which enables the rings to be formed, be a phenomenon of diffraction, or of scattering in the strict sense of